INDEX

Explanations

mentions or references to anonymity or anonymous individuals

oai_token-act-pair · gpt-3.5-turbo

references to anonymity or anonymous entities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

NEY

-0.90

asters

-0.77

akings

-0.74

united

-0.73

forces

-0.73

neys

-0.72

aster

-0.72

PRES

-0.72

FK

-0.72

Eng

-0.72

POSITIVE LOGITS

 anonymous

1.25

 anonymity

1.06

 informant

0.94

 anonymously

0.93

otine

0.90

onym

0.90

onymous

0.90

 bystand

0.90

 anonym

0.87

 whistleblower

0.86

Activations Density 0.010%

mentions or references to anonymity or anonymous individuals

references to anonymity or anonymous entities

No Comments

No Known Activations